Directory UMM :Data Elmu:jurnal:A:Accounting, Organizations and Society:Vol24.Issue1.Jan1999:

Accounting, Organizations and Society 24 (1999) 31±55

The ¯exible ®rm: strategies for a subcontractor's
management control
Jan Mouritsen
Copenhagen Business School, Informatics and Management Accounting, Howitzvej 60, DK-2000, Frederiksberg, Denmark

Abstract
A ``¯exible ®rm'' is one which orients itself towards customers, new technology, lateral organisational arrangements
and innovation. It is the ``new organisation'' where the customers and empowered employeesÐrather than organisational bureaucracy and capital marketsÐare said to govern the ®rm. BusinessPrint is a ®rm committed to ¯exibility.
However, what ¯exibility is, how it is achieved and what its e€ect is on the ®rm's pro®tability are dicult to resolve as
¯exibility may con¯ict with productivity. In BusinessPrint, ¯exibility is debated against two modes of management
control: one is the ``virtual organisation'' and the other is the ``political organisation''. The former is predicated upon
the possibility to inscribe not only the ®rm's internal production processes but also its relations to customers and
subcontractors in an information system and to let that inscription organise inter-organisational spaces of ¯ows. In
contrast, the latter is mobilised via labour processes in local places designed to motivate workers to show unconditional
adaptability and improvisation in production activities. # 1998 Elsevier Science Ltd. All rights reserved.

Flexibility and innovation are often presented as
indispensable to competitiveness (Stalk & Hout,
1990) where scope rather than scale, customerorientation and a concern for quality present an

image of a company committed to serving its customers and responding to the speci®c wishes of the
market. Concerns about ¯exibility and innovation
make companies pay more attention to customers
than to products, more attention to customisation
than to mass production, and more attention to
variation than to repetition (Goldhar & Jelinik,
1983; Milgrom & Roberts, 1990, 1995). But what
are ¯exibility and innovation? How do companies
decide just how ¯exible they wish to be? Flexibility
and innovation may stand out as justi®cations for
a ``plant with a future'' (Miller & O'Leary, 1993,
1994), but how is such a programme transformed to speci®c corporate strategies? They may

also justify a ``new commercial agenda'' (Munro &
Hatherley, 1993) by which lateral relations with
customers are supported through empowerment of
subordinates, but how do management control
strategies achieve this?
Such questions were relevant to managers i
BusinessPrint, a small and arguably ¯exible and

innovative ®rm that sold complex machines to the
printing industry. However, while ¯exibility and
innovation were approved in principle in the ®rm
and generally accepted to be part of its compet
tive capacity, what they referred to and just how
¯exible and innovative it was necessary to be, was
not easy to suggest. In BusinessPrint, manager
debated ¯exibility and innovation in a struggle
over management control strategies. Although all
managers accepted that ¯exibility and innovation
were important and possibly a key to its competitiv

0361-3682/98/$Ðsee front matter # 1998 Elsevier Science Ltd. All rights reserved.
PII: S0361 -3 682(97)00059 -7

32

J. Mouritsen/Accounting, Organizations and Society 24 (1999) 31±55

advantage, why and how it was so was debated

against at least two strategies for management
control: accounting control in a ``virtual organisation'' and production control in a ``political organisation'', which were drawn upon as justi®cations
and explanations in managers' attempts to control
¯exibility and innovation.
Aspects of a potential accounting inscription
(typically drawn upon by the CEO) were mobilised to represent the ®rm's situation as lack of
planning and absence of appropriate information
systems. Here, management control was a paperpractice that drew spaces of ¯ows internal and
external to the ®rm together in an integrated
representation measuring plans vs actual results. It
presented the image of a virtual organisation that
existed as information-¯ows in an informational
mode of management. It was not only, however,
that information was available and the ®rm was
redesigned to allow intervention at a distance; it also
was so that information set aspirations for performance through plans and budgets. It was a normative
inscription that connected the past (``aspirations''
and ``goals'') with the present (``results'' and
``performance'') and the future (actions to close
gaps between ``aspirations'' and ``results'').

In contrast, aspects of a politics in production
(often drawn upon by production, sales, and
shipping people) were employed to establish coordination as communication and talk between
co-present people. For the production people,
management control was a ``hands-on'' practice
which separated organisational as well as interorganisational spaces and treated them individually as places for industrial relation issues to be
negotiated in real time. It presented an image of a
political organisation which existed as struggles
between actors and classes. It was not that information was not produced and used here but rather
that it was used di€erently compared with the
``virtual organisation'' since it was drawn upon
much more selectively and was not seen as a complete representation of a ®rm's concerns. In this
situation, management control was executed
through management by contact rather than via
management at a distance.
In BusinessPrint, the management control
agenda was mobilised against what others would

term a ``new commercial agenda'' (Munro &
Hatherley, 1993) and a ``plant with a future''

(Miller & O'Leary, 1994). These address the problems of being ¯exible driven by the custome
The plant with a future and the new commercial
agenda focus on the lateral ordering of ¯ows from
supplier through production to customers, and
they pay extraordinary attention to the well-being
of the customer (Stalk & Hout, 1990; Galbraith &
Lawler, 1993; Galbraith, 1994; Jenkins, 1997;
Wortmann et al., 1997). To do this, ®rms reorganise their production spaces to support throughput and ¯ows of goods and introduce information
technology to underscore a lateral ordering of the
®rm. The plant with a future and the new commercial agenda empower supervisors and workers
and make them directly accountable to each other
and to the customers. The hierarchical ordering o
the ®rm is transformed into a lateral one, or a
least the ``governable person'' is supplemented
with the ``governable process'' across the plant's
spaces of ¯ows (Miller & O'Leary, 1994, p. 41)
which emphasises throughputÐa set of interdependent material and information ¯ows organised in and across organisationsÐand heralds
¯exibility, productivity, and innovation. Managerial technologies such as cost drivers (Cooper &
Kaplan, 1991), business process re-engineering
(Hammer & Champy, 1993), total quality management (Johnson, 1992), empowerment (Bickham

1996; Phelps, 1997), customer-orientation (Galbraith, 1994), and lateral organisational ¯ows
(Galbraith & Lawler, 1993) all in their ways justify
an image of a ®rm which through ¯exibility takes
its commitments to the market seriously (Schoenberger, 1986; Stalk & Hout, 1990). The commi
ments, moreover, to the market also raise
concerns about communication and information
that tie customers and the various production cells
together (Miller & O'Leary, 1994) in a networked
organisation beyond the ®rm (Sotto, 1997). Such
``new organisation'' is legitimated by the power of
the customer to govern the ®rm, and it is said to
require limited power of bureaucracy and top
management. To a certain degree, the ``new organisation'' is one where the ®rm is freed from the
in¯uence of capital markets and top management,
and the power to rule organisational conduct

J. Mouritsen/Accounting, Organizations and Society 24 (1999) 31±55

seen to be shifted to customers and empowered
employees. This is also where the contradictions

within in the ``new organisation'' arise. On the one
hand, the ``new organisation'' is presented as one
where the ``hot'' rhythm of face-to-face communication and improvisation governs its conduct;
but on the other hand, it is also envisaged to be a
network within and beyond the ®rm co-ordinated
by a ``cold'' virtual reality of information and
communication technology.
The ``new organisation'' is thus a heterogeneous
constellation of issues which bear on di€erent
principles of government. In such a situation,
accounting is intertwined with production and
marketing in a constellation of issues ranging from
product costing via quality, ¯exibility, innovation
and productivity to principles of production management and of customer service. It is in this
matrix of issues that companies, such as BusinessPrint, have to orient themselves. What form of
¯exibility could and should they adopt? Could and
should they be innovative, and if so in what way?
Can they empower employees, and could and
should they? Could and should they concentrate
on the customer, and if so, how would they proceed to develop this idea into certain organisational practices?

In BusinessPrint, one concern raised by the
CEO was the status of indirect production costs
which were said to have increased dramatically
over the recent years. This was the accountability
issue which in a sense propelled the debate on
¯exibility and justi®ed his own position to transform management control into a paper-practice.
Indirect costs were not the only managerial problem in BusinessPrint nor were its problems
de®ned merely in terms of costing, but indirect
costs were a medium that enabled him to challenge
the production ethos of the management control
strategy then in place. For the CEO, the indirect
costs constituted a possibility to devise new contours of a management control strategy that
emphasised managerial technologies such as
accounting systems and production control systems which possibly could penetrate production
matters at a distance. Such a ``simple'' quest for
visibility, however, was not translated merely into
a quest for a more ``truthful'' representation of the

33

production process's utilisation of indirect resources. In e€ect, by introducing a concern for indirect
costs, the CEO mobilised a much wider agenda
that more strategically called forth attention to the
role of production processes in BusinessPrint. As
will be shown subsequently, the CEO attempted t
redesign the ®rm to ®t the contribution accounting
system already in place rather than ``merely'' to
create a new and more sophisticated representation of the ®rm. The CEO attempted to make the
®rm amenable to intervention at the distance
(Miller & Rose, 1990) through managerial technologies already in place: he attempted to redesign
the ®rm around a managerial technology which he
already understood. To make do withÐand possibly expand the power ofÐthe contribution
accounting system and the MRP production control system, the CEO wanted to get rid of indirect
costs and transform them into variable costs. This
included a heightened emphasis on subcontracting
parts of the production process just as it involved
stricter agreements with customers about level of
service (Gietzman, 1996). Subcontractors, customers and the internal production process were t
be inscribed into an informational form of management control in an integrated, virtual organ
sation beyond the formal boundaries of

BusinessPrint. This contrasted with the management control strategy in place which heralde
¯exibility and innovation as practices tied to
locality where people could see and talk to each
other; and where management control was based
on direct, co-present relations between managers,
supervisors and workers in political struggles
about control at the point of production. In this
management control strategy, the impetus to
manufacture products depended on the ability of
management to make workers work ¯exibl
according to the ``whims'' of the market as customer±orientation took it into an extreme production organisation where the ability to improvise
was the driver for quick throughput. Improvisation was the e€ect of managers' ability to persuade
workers to adapt totally to the immediate production requirements laid out by demand in the
form of shipping-orders.
This theme is in some respects similar t
Latour's (1987) point that the translation between

34

J. Mouritsen/Accounting, Organizations and Society 24 (1999) 31±55

speci®c ideas (management control procedures)
and general ideas (management control strategies)
is a network e€ect. The translation between (general) concerns for pro®tability and (speci®c) organisational action may be seen as a process by
which problems and solutions are invented,
assembled and fabricated (Bloom®eld et al., 1992;
Preston et al., 1992; Chua, 1995). As mentioned,
in Business Print, two competing translations
between pro®tability and organisational action
debated how ¯exibility, productivity, and innovation would lead to commercial success. This theme
is not, however, merely a story about the ways a
particular managerial technology could be transformed through ingenious machiavellian, micropolitical strategies (Crozier & Friedberg, 1980;
Law, 1994). It is also a story about the transformation of resistance and struggleÐand thus relations of dependence and autonomyÐin modern
industrial systems where managers attempt to
redesign organisational activities and inter-organisational relations to curtail political risk in
labour processes bothÐas will be illustratedÐ
between managers and workers, and between
owner-managers and employed managers. Moreover, it is a story about the transformations of
space enabled by the prospect of an informational
mode of management (Castells, 1989) by which
control of production at a distance would gain
primacy over control in production in situations of
co-presence. This transformation directs attention
to not only the legal boundaries of companies, but
also to the spaces of ¯ows by which products and
services move between them. The boundaries
between ®rms become blurred since what counted
as ``internal'' and ``external'' can be problematised
via debates on their ``core competencies'' (Hamel
& Pralahad, 1994) and brought about via decisions on technology, customers, subcontractors
and management controls. The ``space'' of the
®rm can be rendered negotiable, and its control
mechanisms may be extended via information
systems so that what is ``external'' may be represented directly in ``internal'' standard costing and
budgetary systems. Such inter-organisational
devicesÐderived, for instance, from the standard
costs of subcontractors' intermediate goodsÐhelp
some ®rms control other ®rms' activities since the

``command over spaces and times is a crucial element in any search for pro®t'' (Harvey, 1989, p.
226) whereby the space of social arrangements
involves a regionalisation that constructs distinctions between centres and peripheries in terms o
power and in¯uence (Giddens, 1984; Lash & Urry,
1994). Therefore, the command of space involves
command over others and their things; and the
command over others and their things is materialised via managerial technologies that give social
interaction a certain form. As a consequence,
managerial technologies, that are intermediaries
(Law, 1994), facilities (Giddens, 1984) or techno
ogies (Latour, 1991), help to mobilise, support and
bring systems of management control about; and
di€erent managerial technologies, or di€erent
constellations of managerial technologies, can
support a ®rm's management control mechanisms
di€erently. In turn, they may leverage organisational and the inter-organisational spaces di€erently for instance through either a localised political
management or a globalised virtual management.
The paper illustrates how BusinessPrint debated
the relationship between management contro
cost, ¯exibility, innovation, and customer-orientation and is organised as follows. The next section
presents BusinessPrint's production and management control systems. Part of this includes a presentation of the focal issue in the ®rm, namely the
costs and bene®ts of ¯exibility and innovation.
Following this, another section analyses the two
management control strategies that competed for
signi®cance in BusinessPrint. In doing so, both
presented solutions, albeit opposing, to the issue
of ¯exibility. These solutions were designe
according to management control strategies that
attempted to seize and de®ne the customer, technology, subcontractors and workers di€erently.
These four elements acted as obligatory passage
that the two possible management control strategies had to be able to account for as they formed
the area of translation between the overall quest
for ¯exibility and the particular form of its realisation. The impact of management control on
BusinessPrint's boundaries is analysed more in the
last section of the paper which debates the way the
product was located in an inter-organisational
spaces of ¯ows.

J. Mouritsen/Accounting, Organizations and Society 24 (1999) 31±55

1. The ¯exible ®rm
BusinessPrint employed, when the research was
conducted, about 320 people and was medium±
sized by Danish standards. Some years ago it had
been through a management buy-out, and cash
was still scarce. BusinessPrint was pro®table,
though, and there was no sense of any threatening
®nancial crisis. Throughout its history, it had
emphasised quality, reduction of inventory, and
increased product and process development and
had a reputation for experimenting with new ideas
particularly in the production management area.
The case study was carried out as an ethnography. BusinessPrint was visited over an extended period (about a year) although the main
interviews were all conducted within the ®rst 2
months. In all, about 30 interviews were conducted primarily with managers at all levels from
CEO to foremen and supervisors. This included
managers from all aspects of the ®rm's departments ranging from RandD over production via
logistics and shipping to sales and marketing and
again to accounting and top management. The top
management (CEO) and the chief ®nancial ocer
(CFO) were, along with a manager of a product
division oriented to designing a new product line
not included in the study, owners of the ®rm,
while in particular the production manager but
also the sales manager and the logistics manager
were employed managers. In addition to interviews, participant observation was carried out as
access was allowed to meetings and oce space
was provided in the ®rm so that presence was
there even on days when neither interviews nor
meetings were scheduled. In all, the possibility to
get to understand a good proportion of BusinessPrint life was there . One omission, however,
worth not only of mentioning but indeed of
drawing to attention is that contacts with workers
and unions were minimal. This access was simply
not allowed, but labour process issues did seem to
be important in this ®rm, as it turned out in the
course of the research, because they were part of
managers' account of their agendas. The analysis
of BusinessPrint is not a labour process approach
per se but the language of labour processes was
used by managers to indicate an orientation to

35

management focusing on a certain oral, communications±based form of intervention and control.
As will be clearer in the following sections, the
analysis of this mode of management control
crafted in the text not primarily as a struggle
between workers and managers, but rather as
piece of politics between managers in the construction of managerial agendas and priorities.
Access was thus both generous (vis-aÁ-vis management activities) and constrained (vis-aÁ-vis
workers and unions) at the same time. This is a
obvious limitation, but even the con®dence of the
empirical material about managers should not be
carried too far. As Law (1994) points out, even in
an ethnography where researcher(s) may be present all the time, one always wonders whether
indeed the most essential bits of transaction, communication and decision making were actually
observed. The question is how we can be satis®ed
that important events have not happened in
place where the researcher(s) was not present? The
answer is that we can not satisfy ourselves on this
point. What has been done in this case study is to
attempt to construct a story, a narrative, which in
coherent form adds the pieces of evidence together in a stream of action and explanation which
can be told and communicated (Boland &
Schultze, 1996; Charniawska-Joerges, 1997). It
therefore perhaps less the individual piece of evidence that counts but the interrelations between
various bits of data put into a stream of a storyline which is held together not only by the sheer
amount of data but also by the logic that ties various items of evidence together.
Why BusinessPrint, then? This ®rm was
approached for its reputation to, on the one hand,
innovate drastically and continually in products
and production processes, and on the other hand,
for its reputation to be highly customer orientated.
It was committed to ¯exibility, the rumour said. It,
among other things, was highly interested in having Ph.D. students and master students around
particularly in the technological area. This
involved not only attention to new materials and
product development, but also an attempt to
introduce modern manufacturing conditions, and
BusinessPrint was one of the ®rst ®rms in Denmark to experiment with self-governing work

36

J. Mouritsen/Accounting, Organizations and Society 24 (1999) 31±55

groups, just as it paid attention to broad just-intime related concerns such as TQM, continuous
improvement and some measure of humanistic
industrial relations programs. These were organisational innovations connected to the manufacturing area; it was less customary to look at the
®rm's managerial area. So, BusinessPrint was
rumoured to be ¯exible, innovative and customer±
orientated, it seemingly had a policy to introduce
modern manufacturing conditions, and it thus
seemed appropriate for a study concerning the
managerial technologies that make ¯exibility work.
The case study was designed to understand
management control and the technologies that
support it in what was thought to be a ¯exible
®rm. It was designed to contextualise managerial
technologies and debate their position in the ®rm:
How does accounting control and production
control interrelate? How does marketing and production refer to each other? How are production
technologies and managerial technologies mobilised? Such general questions were traced and certain ``obligatory passages'' (Callon, 1986; Latour,
1987) were identi®ed which characterised broadly
the problems which any management control
strategy had to be able to answer in order for it to
be deemed appropriate in BusinessPrint. Important obligatory passages were here particularly
``the customer'' and in addition ``subcontractor'',
``technology'', and ``workers'', all of which had a
representational space in the management control
procedures.
1.1. Products, processes, management control
systems
BusinessPrint produced a high-tech productline and sold most of its products (about
95%) through a large international ®rm, ``the customer'', more speci®cally through its subsidiaries.1
It produced about 6,500 units per year (about
35±40% of the world market2), each comprising
2,000±3,000 components. The recent years had

1
In a world-market of only 20,000 units, distribution and
service costs could turn out to be excessive if sales was not
carried out through the customer's net of subsidiaries.

witnessed an increase in the number of products
and variants to accommodate the customer's
wishes.
The production process was simple consisting of
four sequentially interdependent processes: In the
®rst process, metals for the cabinet of the product
were cut into pieces which were then welded together in appropriate forms. The second process was
a painting activity, where the components and
cabinets were coloured and lacquered. Following
this, the third process was an assembly activity
where the ®nished products were put together
from the cabinets produced in previous production processes and from purchased parts such as
electronics and lenses. The last process was shipping where the ®nished goods were put into containers to be sent to customers across the world.
Although capacity was scarce in all processes,
painting was particularly important in this respect.
It was not possible to rush drying the paint, just as
there was set-up time involved in changing from
one colour to another (a half day set-up time, and
2 days drying before the part could be transferred
to assembly). The painting process constituted
separation or barrier between two regions within
the ®rm that were governed by two di€erent production planning mechanisms. Cutting and welding activities were planned for through a push
system devised by a Material Requirement Planning System (MRP) and thus organised via
expected sales in the marketplace. The MRP system broke planned ®nal production volume into
purchases and production orders (Mackey &
Thomas, 1995). The system coupled production
and purchasing in an integrated plan which was
important not least due to the sheer number of
components in the product and to the long time
horizon of the whole production process. The lead
time for one essential component (which is bought
from the major customer) was 8 to 9 months,
while other components (especially electronics)

2
This market was in decline because the technology was
becoming obsolete. However, BusinessPrint expected to be able
to continue producing the same number of units for some time
still as it attempted to capture market-shares. It also diversi®ed
into another technology although in a di€erent newly established ®rm.

J. Mouritsen/Accounting, Organizations and Society 24 (1999) 31±55

had a lead-time of 3 to 4 months. Throughput
time for the whole production process was 31±35
days, although the supervisors could speed the
process up and reduce it.3
Planning in the cutting and welding processes
was ®nely tuned as it detailed the production process minutely from expected sales and orders in
the marketplace. This detail in planning was dicult in assembly because shipping directly controlled the sequence in which orders were taken
out of the factory. In assembly, therefore, there
was more of a pull system, a version of a just-intime system, more than a push systems in¯uenced
directly by the MRP systems' plans. Customers
put their orders directly with shipping which sent
containers around the world. This often disregarded the MRP system's plan because the container space was an important constraint on the
movement of products between BusinessPrint and
its customers. Therefore, shipping often dismantled the MRP system's plan and sought to ®ll
a container and thus complied with the customer
orders that came in every day.
Although the MRP system did consider assembly processes in its proposed plans, these plans
were rarely, if ever, realised in assembly and in
shipping. Due to the long production process in
the painting department, cutting and welding
activities were only loosely coupled with, and thus
to a degree separated from, assembly and shipping
activities. The painting department made tight
coupling dicult as it could not quickly redirect its production vis-aÁ-vis actual demand, and
therefore it was a time±space barrier that separated the plan from the customer.
The MRP system was in line with Business
Print's budgetary control mechanisms based on a
contribution accounting system. The MRP system
focused on standard costs just like the budgetary
control system paid attention to variable standard
costs. In addition there was an increasing awareness of the unallocated pool of indirect costs, and
non-®nancial measures such as throughput time,
set-up time and inventory levels had been in place
3
It was not easy to get an answer to how much foremen can
decrease throughput time. As will be shown subsequently, this
was part of company politics to be able to do this.

37

for years. The pool of indirect cost amounted to
about 35% of total cost and was identi®ed as
problem by managers, particularly the CEO but
also the CFO in BusinessPrint. Its contribution
accounting system did not allocate ®xed costs to
products or product-lines as this was considered
arbitrary, and it increasingly was a mystery to
managers what the rationale of the pool of ®xed
cost was, not least since it grew continuously.
Although there was agreement among manager
that indirect costs did pose a problem, there was
little agreement about what to do about them.
One possibility was to get rid of the ®xed costs
by exporting parts of the production process
through subcontracting. In this way, some, if not
all, indirect costs would be transformed to variable costs and suit the contribution accounting
system that would enable managers to understand
costs as a direct function of the level of production. In this sense indirect costs would be managed
by exporting them and getting them back as variable costs. This position was held by the CEO who
suggested that it would easier for him to engage
more directly in the mobilisation of Business Print
if he was able more systematically to intervene i
questions of production organisation via an access
to productivity through constructing the ®rm's
®nancial structure so that comparison between
standard and actual would be a direct inscription
of eciency.
Another possibility was to manage the indirect
costs directly. Activity Based Costing was proposed by the factory manager to be such a solution. In this way, the indirect costs would still be
signi®cant, but they could be related or traced to
products or processes.4 It was doubtful, though,
just to what extent it would be possible to control
indirect costs even if they could be traced to products because there was a limit to the possible
actions that could be made in reorganising the

4
Although the possibility of Activity Based Costing was
known in BusinessPrint, it was ignored by the CFOÐbut not
by the production people!Ðas it was argued to be possibly
arbitrary and certainly costly to operate. BusinesPrint's
accounting system adopted a contribution principle which
emphasised product-lines' contribution towards indirect cost
and pro®ts. It did little, however, to ``explain'' indirect costs.

38

J. Mouritsen/Accounting, Organizations and Society 24 (1999) 31±55

pool of indirect costs because of the ®rm's limited
size. In spite hereof, the production manager
suggested that a more detailed inscription of the
relationship between indirect costs and production
activities would be important for the general
awareness of the production processes.
These two possibilities were not mobilised
merely as technical solutions to problems of
insight and visibility but they represented fundamental di€erences in preferred management control strategies. The preference for a management
control strategy was not predicated only on the
prospect of a ``truthful'' representation of a production process; it was also predicated on the type
of decisions each manager would be able to execute. It depended on the knowledge and possible
courses of action that could be put in motion. This
is why the CEO attempted to problematise BusinessPrint's problems as lack of knowledge of the
production process which could be recti®ed if just
the production process could be rendered amenable to control in the form of an inscription that
would be able not only to ``describe'' the state of
a€airs; it would also be possible to evaluate the
e€ectiveness of production in the comparisons
between expectations and realised results. In contrast, the factory manager was able to use information more selectively because he was able to
understand to production process directly. He, as
will be illustrated more in subsequent sections, was
able to ``feel'' the e€ectiveness of production
processes and did not need an information system to point substandard activities out. He did,
though, use information but much more selectively as a ``library'' of indices that could be
drawn upon on request: In some periods he
made scrap the important measure, in other
periods the inventory list was prioritised, and in
yet other periods he called for special attention
to quality measures. There was no single representation of the a€airs of the production process. It changed over time and was interpreted
against his ``feelings'' for the state of a€airs.
The problematisation of indirect costs was thus
not merely an ``innocent'' attempt to create additional visibility through a more ``truthful'' representation of production processes. This
problematisation represented a more fundamental

debate on the proper form of management control
in BusinessPrint. This is also the reason why the
issue of indirect costs could not be con®ned to
indirect costs. They also signi®ed a further questioning of strategies concerning ¯exibility, productivity, and innovation. Before this is illustrated
speci®cally, however, it is necessary to delve a bit
more into how ¯exibility was an issue in Business
Print.
1.2. Flexibility as an issue
The CEO and the CFO were critical of Bus
nessPrint's high degree of ¯exibility:
Today, production is managed from the perspective of ¯exibility, throughput and inventory. We don't manage capacity, and we
don't know the costs of our ¯exibility (CEO).
It is not only that the CEO expressed a concern
with the ¯exibility o€ered by BusinessPrint. Also,
and more importantly for the present discussion,
the concern with indirect costs enabled a new
problematisation of BusinessPrint's situation.
Indirect costs represented more than reduced
pro®tability as they also gave impetus to a possible rede®nition of what ¯exibility should be, what
it meant to be innovative, and consequently t
what it meant to serve the customer.
Fig. 1 presents one problematisation of ¯exibility and indirect costs in BusinessPrint. This
presentation is worked up from the empirical evidence and is a more coherent argument than any
speci®c actor suggested. According to this ®gure,
BusinessPrint was ¯exible in as far as it produced
many kinds of products and several variants.
Flexibility could be an asset as it served the customer, but it could also be a liability as it required
resources (particularly of the indirect kind).
Fig. 1 illustrates possible costs and bene®ts of
BusinessPrint's ¯exibility. The relationshi
between costs and bene®ts was, however, not easy
to untangle in the ®rm which also suggests, in turn,
that it is possible to problematise the directions of
the ``causation'' indicated by the arrows shown in
the ®gure. However, the ®gure illustrates a simple
and initial presentation of the issues that were
mobilised in a debate that connected costs and

J. Mouritsen/Accounting, Organizations and Society 24 (1999) 31±55

¯exibility in numerous, and often surprising, ways.
While indirect costs spurred and fuelled a debate
on the appropriate degree of ¯exibility, their association with ¯exibility was a complex and often a
hypothetical one as it was dicult, if not in principle impossible, to produce an accounting calculation which would sort out and evaluate the pros
and cons of ¯exibility.5 Although accounting performance was drawn upon to produce and frame a
debate concerning ¯exibility, it could not de®ne
the desirable level of ¯exibility and it merely
enabled a debate about its pros and cons which is
illustrated in ®gure one. This debate connected
indirect costs and ¯exibility but did little more
than hint at possible consequences of di€erent
modes and levels of ¯exibility.
An important piece of background information to Fig. 1 is that BusinessPrint was a subcontractor to largely one customer that accounted
for 95% of its revenues6 and was large as it had
revenues more than 10 times BusinessPrint's revenues. It was a subcontractor in a special sense
since it did not deliver parts to the customer's
production process but manufactured products
that were complementary to the customer's product. BusinessPrint's product line was a set of
machines that were required to utilise the customer's product line. The customer did have the
opportunity to change supplier, but it did not do
so.7 BusinessPrint's product lines leveraged the
customer's product lines, and the customer was
highly interested in BusinessPrint's product-development activities as these also in¯uenced their
customers' demand for its products. In addition, a
proportion of the components used in BusinessPrint's production process was supplied by the
customer.
The customer that distributed BusinessPrint's
products through its world-wide net of subsidiaries
was large compared with BusinessPrint. Production
and product-development activities were typically
organised around the customer's wishes in a situa5
In principle, this was what Activity Based Costing could be
doing (Cooper & Kaplan, 1987, 1991) since the diversity in
products and the associated complexity in BusinessPrint's support functions could have warranted this. This is what the
production manager said; the CEO would not hear about it.

39

tion where its subsidiaries often placed their orders
late. Business Print did little to prioritise marketing measures by which it could negotiate
other terms of sales with the customer thus
underscoring its image of ¯exibility to the customer. Also, BusinessPrint engaged in heavy
product-development activities oftenÐor typicallyÐin¯uenced by the customer. About 60 out
of 320 total employment were allocated to product development to support the broad portfolio of products.
Both liberal sales-conditions and continuous
product-development in¯uenced BusinessPrint's
indirect costs, as illustrated in Fig. 1. The liberal
conditions of sales were expressed in many rushorders which lead to various kinds of indirect
costs. Firstly, a strain was put on the ``hidden
factory'' (Miller & Vollman, 1985) of planning
balancing and moving. Changing plans was
source of tension because the MRP system was
6
BusinessPrint's relationship to its major customer is a bit
particularistic. BusinessPrint was formed though a management buy out where the present management (the CEO, the
CFO and a third divisional manager but not the production
manager) bought the ®rm from the major customer. Previously,
BusinessPrint was formally a part of the present major customer's organisation. The patterns of communication with the
customer, however, had developed in such a way over the years
that the contacts with the customer were increasingly infrequent and formalised. It would be wrong to suggest that the
management of BusinessPrint was not communicating with the
customer's management team, but it would also be wrong
suggest that this communication was of a broad and varied
nature. There was hardly any discussion of conditions of sales
or of asking the customer to act in consideration of BusinessPrint's needs. Nor was there much discussion about the ®nancial diculties of a high degree of customer-orientation.
Although this might very well have been talked about in the
meetings between BusinessPrint and customer managemen
teams, it was not obvious to the managers in BusinessPrint that
such considerations were legitimate. Management had diculties explaining their concerns to the customer because they
feared that a too aggressive style vis-aÁ-vis the customer would
make it redirect its purchases to the German competitor. This
was their interpretation; an interpretation, as will be shown,
which is open to debate and negotiation. It turned out to be an
element in the struggle between two possible forms of management control to be discussed later in the text.
7
A small part of the customer's demand for products of the
type BusinessPrint could supply was put with BusinessPrint's
main competitor. This was to alert BusinessPrint that it did not
have a monopoly.

40

J. Mouritsen/Accounting, Organizations and Society 24 (1999) 31±55

Fig. 1. A model of cost and ¯exibility.

unable to load new orders both since it took time
and resources to make new plans and since the
redirection of purchasing orders for components
and parts was dicult, if not completely impossible, at short notice, just as rush-orders would not
be able to comply with the planned schedule that
had put a set of production orders in motion. This
produced, it was argued in BusinessPrint, a strain
on the ®rm's planning capacity. Secondly, set-up
costs increased dramatically when new rush-orders
were accepted as workers had to be redirected to
new production lines. This, for example, involved
a ``re-training'' period where workers were to
familiarise themselves with the new product.
Thirdly, this often meant that ®nal assembly had
to be performed twice. It was dicult to ensure
the availability of parts because the production
process itself was time consuming not least
because the painting process was dicult, and
probably impossible, to rush with existing tech-

nology. Supervisors therefore often had to take
®nished products out of ®nished goods inventory,
disassemble them and assemble them anew for
new version of the product.
In addition to these consequences of ¯exibility
in the sales area, innovation in the form of product-development also increased indirect costs in
various ways, as Fig. 1 illustrates. Firstly, the
orientation towards product-development rather
than process-development resulted in complex
products and time consuming assembly work.
Production was, for reasons of motivation of
workers, organised around the individual person
who was to assemble a whole unit of the product
rather than only specialise in a small part of the
whole assembly process. It took about a day to
assemble a product and due to this, inspection was
infrequent and quality problems arose. The failure
rate had been as high as 100% in some product
lines. To mend failures, workers were put o€ the

J. Mouritsen/Accounting, Organizations and Society 24 (1999) 31±55

piece-rate system and on to the hourly-base system
creating additional indirect8 costs for production.
Secondly, low priority to the development of
production processes raised problems as regards
taking new products into production eciently. The
hectic product-development activities were sometimes implemented in production merely as a message on the computer-screen sent from productdevelopment that the product had to be manufactured in a di€erent way. There was little attention
given to design for manufacturability and therefore
the production process was designed less for eciency in throughput than for the possibility to customise products. A craft shop mode rather than an
assembly line organisation appeared to be in place.
Thirdly, because new products were equipped
with new features and new functionality, Business
Print continuously proved its ability to employ up
to date technology. However, the customer saw
this not only as a progressive product programme
but also as an opportunity to ask for similar features and functionality in old products. This created
a demand to continuously upgrade old products
and change them to incorporate new features and
technologies. Consequently, additional productdevelopment of old products was continually
required andÐin a senseÐa product was never
fully developed.
Fig. 1 illustrates how ¯exibility vis-aÁ-vis the
markets was transformed into liberal sales conditions and a high concern for product development,
both of which resulted in an increasing share of
indirect costs. It assembles a set of stories about
¯exibility and costs which together problematise
¯exibility. Obviously, there was no clear cut
answer to the issue of ``optimal ¯exibility'' because
it was not merely a question of minimal cost;
rather, it was a question of minimal opportunity
cost and everybody realised that the hard question
was to what extent ¯exibility actually created
additional revenues? Flexibility was obviously
expensive; but would it be a pro®table investment?
Although BusinessPrint might get premium prices
for its products, why was it con®dent to accept the
customer's drain on its resources? Or, how was it
8

They were de®ned as indirect simply because they were
added to this pool.

41

that BusinessPrint was satis®ed that it had
responded to its the customer with appropriate
measures? The point here is that not all were
satis®ed with this situation! At least, a debate on
the meaning of ¯exibility was created around two
competing explanations of what constituted
appropriate management control.

2. Two forms of management control
A ®rm's strategy is rarely a fully accepted vision
and version of its present and future (Pettigrew,
1985; Mintzberg, 1994). Strategies were emerging
as BusinessPrint's members started to talk about
indirect costs and developed this debate into
concern for its mode of production and its kind of
¯exibility vis-aÁ-vis its customer. As will be elaborated subsequently, managers disagreed on the
wisdom of Business Print's strategy and related
management control procedures. Their versions of
the sources of strategic excellence were based on
varying evaluations of the di€erent management
control practices not only because of professional
disagreements but also because of ambiguities
about how properly to translate between pro®
ability objectives and control practices. In Bus
nessPrint, the quest for ¯exibility was debated
against two versions of what management control
could be about: the ``paper'' version of management control through a possible virtual organisation, and the ``hands-on'' version of management
control via the actual political organisation.
2.1. Paper, visibility, and restructuring: the
possible virtual organisation
One version of the sources of pro®tability
argued that there had to be limits to ¯exibility and
innovation as these were costly and could lead to
poor productivity. Primarily the CEO and the
CFO adopted this position arguing that ¯exibility
be rede®ned so that it could accommodate a more
coupled organisational system than the one presented in Fig. 1. Indeed, Fig. 1 presented for these
managers a problem that had to be managed via
more strict adherence to plans and budgets. Flexibility had to be contained, according to this view,

42

J. Mouritsen/Accounting, Organizations and Society 24 (1999) 31±55

through increased negotiation with the primary
customer and create arrangements that would
condition and reduce the need for ¯exibility. If
sales conditions were restructured to be a little less
anarchistic, the production system would be able
to be run more closely by a plan and a budget.
Consequently, the various organisational entities
would be integrated via a planning procedure
governed by the MRP system. This could reduce
indirect costs because a heightened attention to
eliminating rush-orders would reduce the barriers
to a free ¯ow of components from cutting and
welding via painting to assembly and to shipping
because there would be less exceptions from the
plan. The importance of the barrier introduced by
painting would be reduced.
Furthermore, according to this view, parts of
the production process had to be subcontracted
out to make a larger part of BusinessPrint's
costs variable with volume. This would not only
reduce business risk but also enhance productivity
because di€erent subcontractors could be compared and made to compete with each other on
productivity and price. Last, innovation had to
be curtailed through increased emphasis on project management systems. The focus of management activities had to be to construct and produce
information useful to enhance visibility and to
underscore management as a paper practice. Such
a call for an information technology which enhanced
the possibility to control manufacturing and move
the product from the subcontractors through
BusinessPrint to the customer was constructed as a
response to the ``control problem'' where production
people's lack of interest in controlling ¯exibility
was seen by the CEO to be an attempt to avoid
systematic approaches to production, and therefore
it was indicative that the high cost of ¯exibility
was not warranted. As one planner emphasised:
Really, sales determines production from the
direct in¯ow of orders. If I am to control
anything, I have to control sales. Presently,
we practise an informal JIT system, at least in
the latter part of production.
It is not that the planner did not attempt to control production by planning, it is just that the
exceptions were so frequent that he had diculties

just putting them into the MRP system:
Last year we made thirty plans. Of these
twenty were changes to already existing plans.
The CEO addressed this as follows:
The sales people really in¯uence what goes on
in production. We believe the customer sets
the rules. We play the passive part as we do
not know their sales situation (i.e. forecasts).
The sales manager was blunt about this:
The relationship between production and me
is through the sales budget. I don't concentrate on connections within production.
Production on the one side and salesÐand shippingÐon the other were separated by the budget
and problems concerning the interrelationships
between marketing strategy and production stra
egy were marginalised. The planner argued with
some pain as follows:
Shipping has resisted that orders could be
booked on speci®c units in the production
plan because if som